Performance of Risk Assessment Models for Peripartum Thromboprophylaxis.

OBJECTIVE: There is no consensus on which risk stratification approach to use for thromboprophylaxis in pregnancy, and most available risk assessment models (RAMs) for venous thromboembolism (VTE) events have not been validated in pregnancy. Our objective was to compare the performance of some of the most commonly used VTEs RAMs in our patient population in the peripartum period.

STUDY DESIGN: This is a retrospective cohort of women who delivered at our institution in 2015 and 2016. We excluded patients with history of prior or current VTEs or those already receiving anticoagulants. Antepartum, intrapartum, and postpartum records were reviewed. Data were collected on known risk factors for VTEs in order to calculate scores for 3 RAMs on admission for delivery: Padua, Caprini, and Royal College of Obstetricians and Gynaecologists (RCOG). The primary objective was to the estimate the performance of the various RAMs in preventing postpartum VTEs. We calculated the proportion of women who would have been high risk, determined the number of VTEs cases within high- and low-risk categories, as well as calculated the number needed to treat (NNT) for each RAM. We performed analyses using different RAM scores cutoffs, VTEs outcome rates, and effectiveness of anticoagulation to prevent VTEs. The P value <.05 was considered statistically significant.

RESULTS: A total of 6094 women were included. Three women had VTEs for an overall rate of 0.04% (N = 3; 95% confidence interval [CI]: 0.01-0.15). The proportion of women categorized as high risk for VTEs were 62% (95% CI: 61-64) for RCOG, 0.8% (95% CI: 0.6-1.0) for Padua, and 94% (95% CI: 94-95) for Caprini. Of the 3 VTEs, the RCOG model classified 1 woman as high risk and Padua model classified 3 women as high risk; whereas the Caprini did not identify any women as high risk. Assuming 100% effectiveness of thromboprophylaxis, the observed NNT was 3838 using RCOG, not able to calculate using Padua (no VTEs cases occurred in the high-risk group, thus none were prevented), and 1927 using Caprini.

CONCLUSION: The rates of VTEs in pregnancy are very low and the available RAMs do not identify most of them. The RCOG and Caprini RAMs would categorize a large proportion of women as high risk and are associated with high NNTs. The Padua RAM appears to have the lowest NNT but missed all the VTEs in our cohort.